notification-system.svg

Case Study

Webhooks

### Problem

The possibility to send a high number of emails without keeping your own server for email delivery.
Goals and challenges
- Track user interaction with the email (up to 10K notifications per minute)
- Reduce load on 3rd party app that sends high number of emails daily
- Webhook receiver’s downtime shouldn’t affect main application flow
- System has to be scalable and guarantee to deliver within one day, when hook target isn’t available
- Convenient and timely email delivery status reporting and displaying this info on a dashboard
- Service is easy to use and customizable to meet changing requirements
- Availability and comprehensiveness of documentation

### How WE helped

We've implemented the hooks delivery system with Amazon Simple Notification Service (SNS) and proxy server that processes, subscribes, and redirects requests. Little coding was required and responsibility of webhook delivery is assigned to an external service.

### Results

- Email delivery system with the user’s activity tracking feature that processes a high number of events (up to 20K notifications per minute)
- Information about user interaction with the email is captured and stored
- Redelivery solution is available when webhook receiver is unreachable
- System is separated and does not affect main app services
- Solution is easy to scale
- 2 times higher level of throughput'

### Implementation process

To achieve project objectives, we’ve considered several implementation options:
- Naive implementation
- Implementation via SQS queue
- Webhooks delivery SNS notification

We have the following components that are responsible for events in our system:
1. Event consumer
2. Pull queue and decide what we need to do with events.
3. Event storage
4. Store all events in our system.
5. 3rd party app
6. Have to be informed about events related to the app account.

### Native implementation

The simple way to inform a 3rd party API is just to make HTTP request the hook URL directly from the consumer, as displayed on the following diagram:

Pros:
- Easy to implement

Cons:
- Main application flow is affected, thus if a 3rd party app won't respond, we will receive an error in our internal consumer
- Сonsumer cannot be scaled separately from webhooks which are sent to 3rd party app
- Impossible to implement webhooks redelivery correctly

Accordingly, we came to the conclusion that this approach will not give us expected results or satisfy our needs in a comprehensive manner.

Therefore, we’ve started to consider other options.

![text](/src/case_studies/notification_system/diagram_1.png)

### Implementation via SQS Queue

We use a lot of AWS services at the moment and already have a bunch of SQS Messaging Queues. The main idea for the solution it is to use a separate queue for hooks (webhooks) and consumer to process this queue. Take a look at the schema that shows how SQS queue works:
In this manner our main app flow does not depend on the working state of an external app and we are able to implement retries via SQS queue.

The flow of a retry is as follows:
1. Pull webhooks events from queue
2. Send HTTP request with events in the body to 3rd party app
3. If request is completed successfully, then remove message from queue.
4. If request has failed, retry with the specified delay and send a message once again.
5. Set maximum receive count that equals delivery retry count'

![diagram_2](/src/case_studies/notification_system/diagram_2.png)


The flow is displayed on diagram below:
Pros:
- Main application flow is not affected
- Redelivery can be easily implemented
- Independent main event consumer can be scaled

Cons:
- More complex to implement
- Max redelivery delay is 15 minutes (since max delay is 15 minutes in SQS queues), thus we can’t make a logarithmic retry for our messages
- Logic of retries and responsibility for webhook delivery has to be managed manually

This option was way better, however, not all of our primary goals would be met by implementation via SQS queue.'

![diagram_3](/src/case_studies/notification_system/diagram_3.png)

### Webhooks delivery by SNS Notification

Another way to implement webhooks delivery is by using AWS SNS. The service was developed for delivering notifications and can send them via HTTP, the same way as webhooks do. Also, if notification delivery has failed, SNS is responsible for retry delivery, and we can use linear or logarithmic function to manage retry frequency.

This solution will give us desirable outcomes; however, it cannot be applied due to these two reasons:

1. SNS send confirmation requests to subscribe endpoint (webhooks destination)
2. SNS provides redundant metadata to delivered messages

The SNS was designed for communication between different components within one system and not for communication among different components of different systems with different owners. Hence, we need a proxy component inside our system that SNS can communicate with

Let's take a look at the scheme of proposed webhooks implementation using SNS:

![diagram_4](/src/case_studies/notification_system/diagram_4.png)

Flow of hooks delivery using SNS:

1. The consumer sends events to a certain SNS topic.
2. We subscribe a 3rd party app to the topic and the URL should contain an endpoint that we need a redirect for (http://proxy.example.com/http%3A%2F%2Fwebhook.endoint%2Ftarget)
P.S.: We’ve encoded the URL with Base64 instead of simple URL encoding.
3. When the proxy receives a request for subscription confirmation, it will send a request to SNS with a proper confirmation token.
4. When the proxy receives a request with a notification (an event, in our case), it will format event data and redirect to a URL that contains the URL path part.
5. All responses from the hooks’ endpoints pass as responses to SNS requests.

Schema of proxy API is shown below:

Pros:

 - Main application flow isn’t affected by the state of the hooks’ receiver application
- The responsibility for message delivery is shifted to SNS.
- Redelivery strategy is already implemented by AWS.
- The simple proxy has good throughput, very lightweight and can be scaled.
- Little code is needed (less code, fewer bugs)

Cons:
- The proxy has to be developed.

![text](/src/case_studies/notification_system/diagram_5.png)

### Solution

After considering each implementation option in terms of benefit it will bring, and required resources for its execution, we’ve chosen to go with the webhooks delivery by SNS notification. With SNS and the simple proxy server, we’ve implemented the hooks delivery system with minimum coding involved and shifted responsibility of webhook delivery to an external service. As a result, we are able to send about 20,000 events per minute without any scaling, which is even higher compared to the planned 10,000 events per minute. The major bottleneck is the proxy, but we designed this one for scaling so throughput can be easily increased.

### Additional info

We wrote our proxy using Node.js platform. The proxy hosted on EC2 T2 small instance and can process about 4,000 concurrent requests with throughput of about 400 req/sec. The proxy doesn’t have any interaction with the database since all data for redirects are passed in URLs. This gives us an option to do a horizontal scaling. If we place our proxy behind a load balancer, then we will be able to add more server instances and as a result obtain two times higher throughput.

Vertical scaling won’t give us the same effect. When we scale node processes within one server instance, we can get about 30% higher throughput for each additional processor kernel. Therefore, horizontal scaling is a more suitable and effective approach.

We don’t need to worry about scaling SNS because AWS does it automatically, we only have to send the proper count of messages.


Creating webhooks notification system

s3-file-upload.svg

Amazon S3

### Problem

Users could not upload all their products to the marketplace because most of them were files that are larger than 200Mb or were video files. Both were not supported for upload; there was a size limitation for digital files and video files were not supported at all.'

### Goals and challenges

- Allow users to upload files that are larger than 200 Mb
- Enable users to continue file upload after the network error from user side has occurred
- Scan uploaded files for malware

### How Wise Engineering helped

Direct upload of files to Amazon S3 was designed and successfully applied in the existing service (marketplace) within a shorter period of time than the upload system that was used before: half a month compared to 2 months. Moreover, it also enhanced security for downloaded files.

### Results

- Users can upload files up to 500Mb or up to 1 Gb (depending on association with a specific user group)
- New items (Video Files) became available for upload
- S3 server usage decreased the time and resources required for server maintenance compared to the hardware servers that were used before
- Enhanced security of downloaded files

### Background

Existing service allowed users to upload files and attach them to products. However, this service puts a solid limit on the size of files that are selected for upload. As a result, it was not possible to upload a file that is larger than 200 Mb. Also, it worked in a way that when a file upload failed (due to the network error), the user was forced to start the entire process all over again. A major problem is that many users have files that are larger than 200Mb. Consequently, there was a lot of requests to upload such types of files. Besides, the NGINX module which was used for file upload was very old and unsupportable.

### Implementation process

Old upload schema

Previously, when the user uploaded a file, we used an old NGINX module. This NGINX module simply read all the content of the file and stored it on our server in a temporary directory with a temporary file name. The following problems emerged:

1. This mechanism read the whole file within one operation.This approach doesn’t work correctly for the big files and caused a significant load on our servers.
2. The unique temporary file name was not, in fact, unique.
Sometimes, when we had lot of users on the website, we had conflicts when two or more users uploaded files with the exact same temporary file name.
3. Our upload progress bar displayed the upload process only.We could not add a progress bar for virus check, consequently users just saw a stalled progress bar while we checked the file for viruses. Also, we generated preview files, which were a reduced version of the original user’s files; because of this, users could not monitor the progression of the upload process.
4. After upload, we stored all files on our own data servers.It required for us more attention and resources to maintain them. Moreover, it made our codebase more complicated since we had to have some extra fields in the database in table item properties; we called it mask of servers, which contains information where a file for a specified item was stored. It increased complexity when we needed to select all items which stored files on specific data servers only.
It required for us more attention and resources to maintain them. Moreover, it made our codebase more complicated since we had to have some extra fields in the database in table item properties; we called it mask of servers, which contains information where a file for a specified item was stored. It increased complexity when we needed to select all items which stored files on specific data servers only.
5. This upload process was not really secure:
- If someone hacked our system, he could catch and replace files
- We checked files for viruses by Cron after upload was completed. During the time between upload and check, an infected file may corrupt our system and harm other users who later bought the product and downloaded the infected file.

![diagram_1](/src/case_studies/file_upload/diagram_1.png)

### Solution

New upload schema

We decided to upload files directly to S3 and no longer store them on our servers. Also, we’ve chosen to use a 3rd party JavaScript library to upload files to S3 directly instead of creating our own library. We selected [https://github.com/TTLabs/EvaporateJS.](https://github.com/TTLabs/EvaporateJS. )This library has already implemented and tested features for all known web browsers:


- Secured direct S3 file upload
- Resume failed file upload

Now all files are uploading to S3 into a temporary bucket for each user’s item. Also, we use unique paths and file names. We decided to split files into a smaller parts before upload and run uploads to S3 in parallel threads, each part is signed by our server so it can’t be catched and replaced by someone else. Now when a user selects a file to upload to S3, the following occurs:

1. Evaporate JS uploads a file into a temporary bucket directly. During the upload process, it makes a request to our servers only for signing each part of the file. This is done in order to avoid cases when somebody can catch part of the uploaded file and replace it.

2. After uploading files into the temporary bucket we started to process the files in the following order:
   - Check for viruses:
If a virus was detected, we stop processing the file, move it into a quarantine bucket (it clears automatically after 4 days) and notify the user about this event.
   - Generate preview files
   - Collect information about the file (size, checksum, type of file, etc.)

3. User fills in required text fields and clicks on submit button.
4. Process push job of funneling item into a queue:
When the user wants to create several items, we organize all the processes into a queue. Once the first item is submitted for creation/update, the user is able to select the next one and doesn't have to wait until the create/update process is finished for the first item. Once the item is created/updated, we display a message to inform the user.
5. In the background, worker process takes jobs from the queue and performs the next job in parallel threads:
    - Save item details into database
6. In case the background saving process has failed (network issue, Amazon issue, etc.), the queue process always restarts failed tasks a couple of times.

![diagram_2](/src/case_studies/file_upload/diagram_2.png)

### Results

The new upload process was designed and successfully applied. We gave our users an opportunity to upload files up to 500Mb. For premium users, we’ve added a new feature - opportunity to upload files up to 1 Gb. Also, we decided to allow upload of video files as a new item. This upload process was successfully applied in the existing service (marketplace) within a shorter period of time than the upload system that was used before: half a month compared to 2 months. Using the S3 server also makes it easier to maintain, and takes less time and effort to maintain, compared to the hardware servers that were used before. Moreover, it also reinforced security for downloaded files.


Direct S3 file upload

scheduled-emails.svg

### Why?

Our main target is to develop a system for sending large numbers of emails and then tracking users’ interaction with a received mail (e.g. opens, link clicks, unsubscribes, etc.) Users with large amounts of daily sends prefer use scheduling for sending use.

### Why we need scheduling

Using scheduling, instead of send-immediately flows, gives us several advantages.
- Better control for sending process
- If a user needs to send emails after his workday is finished, then he won’t be able to monitor the sending process and fix them if something goes wrong. With scheduling, users can prepare all emails for sending before the actual send date, and our ESP (Email Service Provider) handles the rest.Less resources required
- If you already have content for emails that you need to send (for example tomorrow), you are able to send the emails gradually. For example, if you need to send 5M emails at 8 PM you can start sending at 10 AM and continue sending in small batches the whole day. We collect the emails from our side and cover the delivery in time.No personal CRON.

If you have a notification email that needs to be sent after some time (for example, a subscription reminder), just send the notification to our ESP and we cover the scheduling. You don’t need to handle that case with personal Cron scripts.

### Problems that we faced when building scheduling

We propose pretty simple API for working with scheduling, but it’s not easy to organize simultaneously sending even 1M emails. We needed to think about scalability and our system in the future. We had to solve several problems.

- Proper storage for scheduled emails
- Limited throughput of the network for one process
- Split scheduled emails that were ready to send for different workers
- Syncing statuses of scheduled emails and avoiding duplication on send

### Proper storage for scheduled emails

We should able to filter “ready for send” emails as fast as possible among large amounts of records that don’t need to send yet. At the time we already used ClickHouse database as storage for our analytics events (unsubscribe, clicks, open, etc.).

Clickhouse is a column-oriented database that is used mostly for analytics. But with the read/write throughput that ClickHouse provides for us we are able to use the solution as scheduling storage.

The most valuable storage requirements for us were the following:
- Local and distributed joins
- Data compression
- Parallel and distributed query execution
- Cross-datacenter replication
- High availability
- SQL support
- Fetch speed not less than 500K email records per second.

### Performance testing

We made some performance tests and got results that completely satisfied our requirements:

- Read speed is 1.2 GB/s
- Write speed is 50 to 200 MB/s
 - Finding 335K scheduled emails on a dataset with the size of 5M takes 0.061 second

As a result, with network latency for transfers, we were able to fetch 3.5M emails per second. That was a very good result for us.

### Limited throughput of the network for one process

We resolved the problem of choosing our data storage and then faced another issue. We used NodeJS as our main platform, the platform, together with SQS queues, didn’t provide the desired performance for us. Then we wrote simple prototypes for testing using Golang and Elixir (we hoped that Erlang with its lightweight threads helps us) we got some throughput increases, but still weren’t reaching the desired results and scalable solutions.

Results of our benchmark’s queue inserting time for 50K messages to queue was as follows:
- NodeJS 1 thread = 95.408 sec
- NodeJS 2 thread = 75.757 sec
- Go = 35.564 sec

Elixir was about 2 times slower than NodeJS with 2 threads and the same kind of problem.

So obviously, we needed to find ways to scale horizontally.


### Split scheduled emails that are ready to send for different workers

Obviously, we needed to make our scheduling scalable and found a way to split all scheduled emails for workers. Workers should have the same role and know nothing about each other. It’s allowing us to run more workers on more servers when we need to process the largest amount of emails.

When workers are fully independent from each other, we decided to create one more component in our system.

The component is named Ranger and is responsible for splitting ranges for our workers. Each range shouldn’t intersect with other ranges.

With Ranger and Clickhouse working together, we created the following architecture for our scheduling feature:
1. Emails send to REST API, and if a parameter with a future send date exists, API puts the email to Messaging Queue
2. Write to the consumer processing the Queue and save all emails inside Clickhouse database
3. Ranger runs query for fetching emails and the query contains the following steps:
		1. Records filtered by send date and each time we fetch all records that are ready in that moment
		2. Filter records by time-created range, we need only those that weren’t affected by the previous step. For that, we fetch only records that have a date greater than biggest date from previous execution
		3. Aggregate results by CreatedTime and COUNT() function
4. Set first fetched date as DateFrom for query
5. Wait for full batch size and set last date as DateTo
6. Update Date to as last processed range
7. Send range to Queue for processing by worker
8. Consumer pulls a Queue with ranges and fetches by CreatedTime emails for sending to user
9. Each email sends to target user


![diagram_1](/src/case_studies/scheduled_emails/diagram_1.png)


### Avoid duplication

AWS SQS service promises not less than one delivery for a Standard Queue. this means that sometimes we can obtain one message on two workers and send our email twice. Since each message can contain multiple emails (our current batch size contains 5K) we need to organize duplicate handling.

For that, we have the Redis database and we store each SQS messageID as a Redis key. The keys have an expiration date that's equal to message visibility timeout in the Queue. When a consumer obtains a new message from SQS, he checks for an existing messageID in Redis. If the message already exists, the consumer skips this one.

### Conclusion

Using ClickHouse and AWS Simple Queue Service, we build a scalable scheduling system for delivering emails. These flows and tools can be applied to any actions that could be scheduled, not just for emails.

At the moment, we can fetch 335K records per second, filtering by ready date takes 0.1 seconds for 5M database set (including network latency), and workers that process the emails can easily scale horizontally for the desired throughput.


How to send large amount of scheduled emails simultaneously

email-generation.svg

Graph API

Elixir

### Problem
Each user from the marketplace of 10 million users has an individual set of products that are recommended for purchase. Actual up-to-date data on products recommended for purchase must be collected, stored, and delivered to each user in a “Resources you may like” (RYML) email. A way to generate RYML emails with the relevant content and send them to users on a regular basis has to be developed.

Goals and challenges
- Gather actual up-to-date lists of products recommended for purchase for each user.
- This data must be stored and then delivered to all subscribed users via “Resources you may like” (RYML) emails on a regular basis.
- Delegate email creating and sending sequence functions to email sending service.
- Maintain business logic of collecting recommended data and providing it to the email sending service.

### How Wise Engineering helped
The Recommendations Email Generation Pipeline solution was developed. It allows us to gather accurate information about products recommended for purchase for each user and store this info afterward for a needed period of time.
Results
- Information about recommended products is being accurately collected.
- Email Pipeline solutions gives an ability to transfer collected information regarding 10 million users to the email sending service within only 5 hours.
- Email sending service gets the information if a particular user has any recommended products, and can easily send that list in email messages on a regular basis.

### Background
There is a marketplace with 10 million users and each user has specific products that are recommended for their purchase. This information must be collected and delivered to the user on a regular basis.'

### Implementation process
To begin with, we focused on how to orchestrate composition of microservices and to queue extracted products’ recommendations data. Next, we’ve evaluated options for transforming this data into a format that is readable for third-party services. Subsequently, we’ve designed a way to load this data into an email sending service.

The following tools and techniques were applied to achieve these goals:
1. PHP script to go over all users to group them and chunk by task type (if it has recommendation or not). After that, assigned user ID chunks into queue.
2. PHP workers process the queue with user’s ID data and retrieves recommendations for each of them.
3. Elixir Graph API application exposes data on recommended products to PHP workers. Next, it processes products’ recommendation data and applies some internal rules and limitations on it.
4. Python REST API application generates product recommendation data for a particular user ID on demand.

### Solution

We came up with the solution and called it “Email Pipeline.” It starts with running PHP script to go over all users and creates a bunch of async tasks. Each task contains information about:
- user IDs that should be processed
- what kind of operation needs to be done

It may happen that some users do not have any recommendations. In this case, we have to clean up their data on the third-party email service. The rest of them should be updated with new recommendations data. This diversification happens on the fly and group users based on the type of task.

The tasks creation process also has a time ramp up feature. Task processing jobs are heavy and time costly operations. To mitigate that, pipeline services run on Kubernetes’ cluster with autoscaling, depending on their load. So, to make sure that autoscaler has enough time to ramp up the cluster size, we create tasks with ramp up time. We increase the rate of tasks being created in 1 hour. After that time, we spawn tasks as fast as we can as all services are scaled and ready to process increased loads created by task processing jobs.

While tasks are being created, workers start to pick them up and process. The process starts with defining the type of task. In case a user doesn’t have any recommendations, it goes straight to a third-party email service and nulls recommendation data for this user. If the user has products recommendations data, then we need to fetch it fir

To fetch this data, we call Elixir Graph API service which represents a middleware between tasks, or in our case, frontend and recommendation storage. Recommended products data itself is stored in a non-precalculated state hold by Python microservice. This service can multiply matrices on the fly and calculate scores for possible product recommendations and return the data to the caller. After Graph API calls and receives data from Python microservice, the data needs to be processed and cleaned up as this service just keeps the list of recommended products and doesn’t know anything about the user. The cleanup process means removing items that were bought by a user already, were wishlisted, added to the cart, or even seen somewhere else on the site. This information is accessible by Graph API so we can easily clean up recommendations we don’t want to be served to the user next time.

We now have information about the set of products recommended to each user in the task, regardless whether it is nulled data or new recommendations. The only step left is to transfer this data to a third-party email sending service. We will use API of the third-party email sending service to perform this.

As a result, the email sending service has the information if a particular user has any recommended products, list of recommended products if there are any, and can easily send that list in email messages on a regular basis.

### Results

The Recommended Products Emails Generation Pipeline gives the ability to transfer collected information about relevant recommended products concerning 10 million users to the email sending service in 5 hours. This allows us to send “Resources You May Like” emails on a daily basis. All tasks that are related to email composing have been delegated to a third party email sending service. This gives us an opportunity to send emails to all the users with one click after the info on recommended products has been imported.


Recommendations email generation pipeline

05_cases-01.svg

Search

# Problem

Nowadays search becomes a primary source of resource discovery. So having a modern search engine that powers it is a vital part of a marketplace or other e-commerce site. E-commerce platforms contain gigabytes of data, usually not well structured and being input by different sellers so searching within it is not a trivial thing.

There are a variety of site search engines and services built on top of them so it is not an easy thing to choose the one that will perfectly fit requirements and be a good platform for further search results optimization and improvements.


# Goals and Challenges
- Choose between SaaS and self-hosted engines
- Flexible search results tuning mechanisms
- Relatively low infrastructure maintenance
- Reach search experience features implemented


# How Wise Engineering Helped
We have created a list of available search engines and services on a market that fit the required goals. Helped to create an evaluation matrix with a list of required search features and tuning capabilities. Filtered candidates to 2-3 vendors and created a plan for further evaluation of search results, integration efforts and further infrastructure maintenance.


# Results
- Created a finalized list of 2-3 potential vendors to evaluate.
- Listed search features to be added with new search vendor.
- Planned further evaluation process, tools and KPIs.
- List regressions and workarounds if current search implementation has some specific use cases.


# Implementation Process

## Evaluation Matrix
First, we created a long list of search engines and services and combined them into a matrix where we’ve put features and capabilities we want to have in a new search engine or provider. Each feature has its weight so it sets priorities and helps to evaluate a final result better. Some engines/providers may have features implemented in a different way and not always meet requirements. In this case, additionally to feature weights we created another feature evaluation dimension with three simple states as “doesn’t meet requirements”, “par” and “exceeds requirements”. So if the feature we want to have in search is important (which is defined in its weight) but has “doesn’t meet requirements” value for it we exclude this search engine/provider from the further evaluation process. At this point, we had an evaluation matrix with only 3 engines and services left that met all the requirements we’ve set on the first step. It was a mix of services and engines to be self-hosted so we needed to create different evaluation and implementation plans for them.

## Sales Calls and Estimation
For services, we scheduled calls with the sales team and went over features we wanted to have, implementation concerns, etc. At this step, it was important to estimate an annual cost of search service according to the features we want, search rates, potential index size, etc. The final price was an important factor in final vendor decision making. For self-hosted engines, we calculated an estimated hardware quantity required to support the search. Also, estimated a potential ops time required to support search infrastructure which also leads to additional cost.

## Demo Search Results Evaluation
There are a few more additional steps we took to help. We created a step by step plan for demo search evaluation. This plan included a mockup of possible search results evaluation tool to run side by side comparison of current search results and demo search setup with new vendors and services. We also created a regression list of features in current search implementation and possible workarounds in new ones. There were some domain-specific search filters and search query expansions that weren’t possible to port to every new search candidate, so we designed a workaround for them to make sure new search candidates didn’t get rid of some search features users may have been used to.

# Results

Evaluation matrix and further vendors evaluation process have been created. We planned tools to be built and KPIs to measure and prepare for a final search engine decision recommendation, regressions, and possible workarounds. A list of new and modern search features has been created, they will be implemented after a new search vendor is chosen and integrated.


Choosing search vendor

search.png

E-commerce

An e-commerce store without a good search is like a grocery store without labels, you'd be left wandering down the aisles of the store with the hope of locating what you are "searching" for by chance. As the store owner, this is bad business as some customers might get jaded before finding their "search", and the outcome of this means no or little business. This can also be applicable to an e-commerce store but replacing the attendants with the search widget or site search option, this illustration's outcome has made the need for a search widget unquestionable.

The efficiency of this widget has surpassed the need for navigation on the website by users but has also become a great tool for the store owner such that, if properly optimized and wielded rightly, it drives sales and increases traffic. The tech surrounding the functionality of the search widget is beyond the generic search present on most websites as this only gives results based exclusively on the search keyword, for example, if the word "shirt" is typed into the search bar, the search results are not expected to go beyond the word "shirt" and we could say this type of search is based strictly on the search keyword and not influenced by the user experience and all other factors.

![Copy of cases-image-1.png](https://wise-site.fra1.digitaloceanspaces.com/Copy_of_cases_image_1_57331e4918.png)

The need to rightly optimize the search, as the effectiveness of the generic search has been questioned. Currently, there are two main options which we will try to describe at a high level. For most CTOs this sounds like "build vs buy" and in many cases, the decision depends not only on what is better but what kind of company you are in: are you a tech-first company or your business first company.

 

## Building an in-house search for e-commerce

 

Building e-commerce search yourself involves you employing the use one of search engines like Elasticsearch, ManticoreSearch or Solr but there are a handful of search features to consider and scope when building e-commerce search, like search types, autosuggest, typo tolerance and many other built in features require a cross-functional team and tools.

These search solutions provide a very flexible search platform. Setting it up can take a month at most, and this relies upon certain factors like the number of products and so on. The need to structure and optimize data for search cannot be left out as it is crucial to the functionality of the solution. The data should be prepared for search: which includes normalizing attribute values, the configuration of basic and advanced features that would improve user experience follows structuring and optimization of data then the ranking of search results based on their relevancies which can be carried out by developing custom ranking algorithms and also specific search functionality based on the specificity of the e-commerce store has to be developed, search functions like autosuggest, search voice assistant likes. And all these have to be pulled together by a team that will dedicate close to 12-15 months for the completion of this search tech.

Pros and cons of using self-hosted search

![Pros and cons of self-hosted search.png](https://wise-site.fra1.digitaloceanspaces.com/Pros_and_cons_of_self_hosted_search_fe3f1f05f1.png)


 

## Adopting 3rd party vendor

 

SaaS companies specialize in specific niches and always optimize their products to get excellent results based on the niche the product specializes in. They have focused on creating an excellent site search that possesses features like text auto-complete, spell checker, natural language processing (NLP), advanced stemming, and more.

Pros and cons of using SaaS

![SaaS.png](https://wise-site.fra1.digitaloceanspaces.com/Saa_S_7026cdb8f0.png)


 

## Where to start?

 

If you are wondering where you should start while doing your research on how to improve search on your site there are a few questions to answer first. Do you know what problem you are solving or what goal you have? If so - great. If not, you have to start from there.
If you have an answer to this question here are a few basic metrics you should track to understand how your search performs and affects sales.

**Conversion Rate**

Conversion rate is arguably the most important metric, it has a significant impact on your site's profitability. When your website search function as well as it should, it becomes a salesman between users and their desired destination pages. That translates to an opportunity for engaging and converting a potential customer.

**Bounce Rate**

The bounce rate is the rate at which new visitors visit your site and immediately click away without doing anything, and you will want to minimize this especially if the bounce rate after the search is high. A high bounce rate can mean several things and common problems include poor design, low usability, or high search time. This is where your potential for improvement indicator is.

**Search Page Speed**

Page Speed is the amount of time that it takes for a web page to load. A page's loading speed is determined by several different factors which include a site's server among others and this issue can be solved by site search providers or the in-house developing team.

**How Deep Search Users Go**

This is as simple as it sounds and just entails how deep search page users will go. A lot of users get jaded after searching the first and second pages, so it's up to the search provider to optimize the search tool to bring results relevant to the keywords up in the search page. Key things you need to keep in mind while working on improving your search.

**Relevancy Ranking**

The task of ensuring that the best-matched results show up first becomes more difficult, ad data grows and negligence of this factor can result in bad search user experience which is not a good thing for your e-commerce store. Constantly improving your data tagging, assessing site search performance, and tweaking your relevance algorithms will be key to search relevancy.

**A/B Testing**

This is an important tool that involves a performance comparison between two versions of a software (a search results page, in this case). A/B testing can be used to determine which search changes to launch after monitoring certain metrics like the ones listed above, then your A/B testing approach will have significant implications for what kinds of changes you need to create, test, and then launch.

**Team Cooperation**

Last but not least building an in-house search for e-commerce definitely starts from a team. Talented engineers, data analysts, and marketers are required to work in an across-functional team together. Good chemistry between team members will ultimately mean the team can move fast toward achieving results.

In conclusion, the pros and cons of both options have to be considered in terms of how it fits the e-commerce store and the budget, among other factors.

Choosing how to develop your e-commerce search, shouldn't be a decision to jump into. Teams have to define what goals they want to achieve by working on search and evaluate all options carefully before making any decision.

Building an in-house search for e-commerce

sales (1).png

The <a href="https://www.retailresearch.org/online-retail.html" target="_blank">center for Retail Research</a> published a report which suggests that e-commerce is the fastest-growing retail market in Europe. With the Coronavirus pandemic keeping millions of local businesses in lockdown, the global e-commerce market is undoubtedly going to <a href="https://www.statista.com/study/42335/ecommerce-report/" target="_blank">surpass 2 trillion dollars</a> by the end of 2020.

Many successful and aspiring entrepreneurs regard e-commerce as one of the most accelerated and efficient systems for reaching successful outcomes online. And to become a successful e-commerce enterprise, business owners will need to do a lot more than just identifying a product niche, set up a website, or advertising to potential customers. Therefore, they will need to improve their search in e-commerce to provide "almost perfect" search queries for potential customers looking for a specific need.

With regards to a recent e-commerce search experience report by The Baymard Institute, you'd be surprised by the outcome. Out of the 50-e-commerce website evaluated, 34 percent of search queries on their website don't usually reveal useful results, and 70 percent of those search engines on those websites do not show relevant product similarities.

![Search Query Support Among Top 50 US E-Commerce Sites (1).png](https://wise-site.fra1.digitaloceanspaces.com/Search_Query_Support_Among_Top_50_US_E_Commerce_Sites_1_e6e15eddce.png)

Whether you manage a small online store or a big e-commerce website, ensuring your e-commerce site search is consistently enhanced to provide frictionless user experience is vital for higher conversion.

 All e-commerce website visitors are unique. Some visitors already know what they want to buy, while others look for entertainment - surfing through the latest trends, product images, description, or checking out the information. Nonetheless, no matter the site visitors, whether browsers or researchers, you can use them to your advantage with your federated search information.

It's important to know that every element on an e-commerce site has a role to play for the success of the site, the same way different shoppers rely on various aspects on the site to shape their user experience. Extensively, online customers can be separated into two transcendent sorts: researchers and browsers.

Researchers are goal-driven shoppers to purchase soon. They already have the idea of what they want but will probably be searching for the most suited option from their search in e-commerce sites.

Researchers may surf through your website multiple times, gathering information before they decide to make a purchase.

On the other hand, browsing shoppers visit an e-commerce site to see what is new, famous, and what's on sale.

They go to their favorite website or a new section to stay up on the latest trends, window-shop for future purchases, or to kill time. Whatever the case may be, having people spend time on your site can be intuitive for your brand. If they are the excellent user experience your site has to offer, they'll be more likely to revisit your site when they are ready to make a purchase. While they are busy getting entertained, you, the business owner has the opportunity to gather valuable information about your visitor, which can help you turn those browsers into buyers.

Another intense research from the Baymard Institute suggests that e-commerce site search is undoubtedly found to be the preferred product finding strategy by shoppers as they perceive it to be faster than category navigation. In this regard, Effective site search plays a direct, quantifiable role in increasing sales revenue, so working towards improving and optimizing your product inventory will directly affect the success of your search in e-commerce.

When on-site search is compared to other metrics, site search is usually ignored. Business owners prefer to pay attention to optimizing personalized emails or landing pages. In contrast, almost all product-focused buyers use on-site search to hasten their buying journey on an e-commerce website. And if your website happens to lack the efficiency and accuracy they want, they will look elsewhere.

Your on-site search plays the role of your sales team. Nonetheless, taking the necessary steps to improve your site search can lead to happier customers who eventually convert to real paying clients.

There are three primary ways shoppers can find products on an e-commerce site:

![Jeans_.png](https://wise-site.fra1.digitaloceanspaces.com/Jeans_5cdcfc2aa2.png)

- Category Browsing: occurs when shoppers surf through categories and narrow down to see the items they are searching for in the subcategories.
- Keyword Search: happens when a potential shopper uses the on-e-commerce site search bar to find products using closely related keywords. The keywords can be short tail or long-tail keywords; internal, competitive, and industry-specific.
- Faceted Search occurs when shoppers refine their search by narrowing the results through search filters. For instance, it can be filtered into length, size, brand, or weight; this type of search filters can be searched on the Amazon website. It will be an excellent opportunity to integrate a solution that will help achieve your search goal - such as enterprise search, or database search engine into your faceted Search to produce more search relevant information to your site visitors.

Several factors contribute to whether shoppers can easily find products through on-site search. These factors make it easier to provide the nearest search queries to visitors, which include synonyms, predictive search, misspellings, and type-ahead. Big websites like Amazon have integrated enterprise search as a service software to ensure their on-site search is performing efficiently.

For instance, Amazon tries to find relevant, accessible, and informative content to provide to its customers. The website uses a sophisticated search algorithm that has one primary objective; bring shoppers closer to relevant products as fast as possible. Basically, the Amazon search algorithm works similarly to google search algorithm. It analyses merely search queries for relevant keywords and matches customers' needs with related products in that niche.

Aside from providing engaging value to its customers, Amazon has access to a considerable amount of data to use for search engine results. This benefit is linked to the fact that Amazon is a self-contained platform, gathering data from many sources and storing them within its big data infrastructures.

Amazon uses the A9 ranking algorithm, whose primary purpose is to deliver a spectacular customer experience, thereby leaving the customer hooked on the website. It achieves this by analyzing different pieces of information to determine the most accurate and valuable product to appear in the search queries. The algorithm usually considers the following criteria to rank the product according to search relevance. They include: keywords, customer reviews, performance history, delivery time, and, most importantly, sales conversion.

 

## Conclusion

 

Whether you are a small or big e-commerce website, the importance of search in e-commerce cannot be overemphasized. Search is relevant and improving search leads to better conversions and engaging shopper experience.

How search affects sales in e-commerce

security (2).png

Security

Global retail e-commerce sales are expected to reach <a href="https://www.statista.com/statistics/288487/forecast-of-global-b2c-e-commerce-growth/" target="_blank">$4.13 trillion</a> this year as online buyers converge on their preferred platforms to buy essentials, luxuries, and everything in between.
As online retail continues to gather speed, e-commerce platforms are being targeted by cybercriminals looking to hack sensitive customer information and scam unsuspecting people.
According to the <a href="https://pdf.ic3.gov/2019_IC3Report.pdf" target="_blank">FBI IC3 2019 Internet Crime Report</a>, global e-commerce merchants reported a loss of over $3.5 billion to cyberattacks. This is why online merchants must secure their e-commerce websites and establish robust security measures to safeguard their data.

So, how can you protect your e-commerce platform and protect customer data?

 

## Know The Areas Of Your Platform To Fortify

 

E-commerce merchants need to keep tabs on several channels, including:

 

### Admin Side Access 

 

Admin portals shouldn't be easily accessible from just about any IP address.

Ensure you lock down cPanels so only permitted IP addresses can access your servers. Next, check if directories have the correct permissions. For example, you'll need to remove all 777 permissions that allow virtually anyone to read, write, and modify files.

Want to take it a step further? Install panel notifications for unknown IP log-in attempts that will alert you and other admins of any suspicious activity.

 

### Get A Reliable VPN

 

Financial transactions over public networks are vulnerable to Man-in-the-Middle (MitM) attacks where malicious users attempt to intercept information being transmitted.

A Virtual Private Network (VPN) offers the best security in situations like these. It encrypts connections to secure offsite servers, eliminating the threat of third-party users who may be trying to intercept your data transmissions.

If you are looking to save costs opt for SSL-based VPNs, these are relatively affordable and still offer similar security specifications.

 

### Infrastructure Security

 

Hackers are employing increasingly sophisticated tools to attack and phish information from online businesses. This means you can't prevent malware attacks by simply installing anti-virus software.

It may be time to invest in powerful application firewalls that let you block communications from unknown domains, and help regulate incoming and outgoing traffic.

Similarly, plugins also help to watch out for and deny unauthorized access to your infrastructure. These tools offer selective permeability that grants permission only to trusted users.
 

### User Data

 

44% of organizations reported <a href="https://www.accenture.com/us-en/insights/security/invest-cyber-resilience" target="_blank">over 500,000 customer records</a> stolen over last year alone.

This is cause for concern as losing personal and financial information of your customers can be the reason for a lawsuit and mar your reputation as a responsible business, discouraging others from visiting your site or buying from you.

This is a liability you want to avoid. As a best practice, you should never store customers' credit card information on your servers. In the event of an attack, this leaves their information unprotected and exposed to misuse.

You might expect clients to get frustrated by re-entering their payment information time and time again. However, they'll be more than understanding of this once you take the time to explain that this is crucial to protecting them from fraud.

Likewise, guide your employees on how to manage sensitive customer information, and ensure they adhere to mandated security protocols, like non-disclosure of users' personal data via email.

 

### Workspace Security

 

Online retailers can (and should) secure their SaaS and cloud-based workspaces against data breaches. Given that your employees access and manage sensitive company data on the cloud, centralizing this information can shield you from data leaks and theft.

To add to that, requiring employees to people to use multifactor authentication can prevent unauthorized access to corporate apps and files - thus strengthening your workspace security.

 

### Personal Training

 

Simply put, one of the safest ways you can avoid malicious attacks and protect your business is to monitor your site for any suspicious activity.

You can leverage monitoring software that tracks user activity in real-time and notifies you of irregular behavior (in the case of cookied users) as well as questionable transactions.

 

## Security Measures To Implement

![cases-image-3.png](https://wise-site.fra1.digitaloceanspaces.com/cases_image_3_6515c8a3d1.png)

 

### Internal Audit

 

Regular internal audits can help you investigate red flags and identify phony profiles. They also highlight potential threats to your online business, such as customer information scrapping, malicious attack footprint, business logic abuse, and more.

Some areas you might want to reviews include:

- Security perimeter
- Consumer activity
- Data integrity

 

### 3rd Party Audit

 

Conduct an inventory of any third-party solutions you run within your business. Continually assess their performance, and if you're not using them, remove their integrations from your systems.

Ideally, you should aim to allow as limited access to your data as possible. If you're not sure where to start, external auditing services can help you assess the value of your inventory - and where you can eliminate threats.

 

### Employees Training

 

More than <a href="https://www.csoonline.com/article/3153707/top-cybersecurity-facts-figures-and-statistics.html" target="_blank">80% of reported security risks</a> are made up of phishing attacks - which are also widely recognized as the top data breach threat to organizations today. It's also one of the easiest ways that malicious users can access your database and misuse both customer and employee data.

If employees can recognize suspicious emails, they'd be less likely to give up sensitive information in the wrong hands. According to Barracuda Networks, spear-phishing emails were <a href="https://blog.barracuda.com/2020/03/26/threat-spotlight-coronavirus-related-phishing/" target="_blank">up by 667% in March alone</a>.

Educate your staff about security best practices and inform them of how to defend themselves against spammy emails and phishing links left in comments on your website. Urge them to alert these IDs to management, who can then take action to identify and eliminate the threat.

 

### Disaster Recovery

 

In the event of a data breach, the best security measure you can take is to make frequent backups of your e-commerce website and mission-critical data. This way, it becomes easier to retrieve information from your hosting provider and restore normalcy to your platform as soon as possible.

Ensure you have a plan in place for managing customer concerns and securing your business during this time, as you'll need to reassure customers you have things under control.

 

### Plan Testing

 

As an added safety precaution, perform regular vulnerability scans across your e-commerce website and assess security gaps that hackers can exploit. This is crucial, especially when you're testing out new features and functionality before updates go live.

 

## Key Takeaway

 

Research states <a href="https://www.globenewswire.com/news-release/2020/04/28/2023282/0/en/Arkose-Labs-Reveals-20-Spike-in-Fraud-as-Digital-Behavior-Shifts-During-COVID-19.html" target="_blank">approximately 445 million cyberattacks</a> have occurred since the start of 2020.

It only takes one successful attempt to access your database - and your entire business is compromised. Protect your brand and your reputation by evaluating your security measures and consider how each can be optimized to fortify your business against external threats.

Ultimately, offering your clients robust security in the face of data threats will help you build their confidence and gain their loyalty.

Cybersecurity best practices for e-commerce merchants

communication_model_main.svg

Management Practices

At Wise, we believe effective communication is the key factor that influences the outcome of any project. Be it collaboration with a client, teamwork on the in-house startup, or the company’s employer brand initiatives. Yet, usually effective communication is hard to measure and even harder to establish.

In this post, we’re reflecting on how communication has transformed in the outsourcing sphere over the last few years. We’ve been developing our clients’ [custom software](https://wise-engineering.com/case-studies/) and launching in-house products for over a decade now and observe some distinct transformations. Let’s discuss communication models in today’s software development outsourcing as well as reasons that forced the changes.

 

## Client-manager communication model

 

Usually, when browsing for an article on communication challenges while outsourcing you’ll read about time differences, cultural peculiarities, and poor language knowledge. These factors are, undoubtedly, important, yet, if you collaborate with an experienced and professional outsourcing team, rarely impact the project outcome that much. What then makes some of the projects succeed and others strive for existence? 

Over the last decade, the approaches to building and launching tech startups have evolved significantly. We no longer launch ideally implemented, bug-free products, plan all the functionality we’ll launch 3 years in advance. Instead, we test hypotheses right from the start, showcase clickable prototypes, get users’ feedback, and proceed with those products or features that get the highest evaluation. 

At the same time, the communication approaches in custom software development companies often remained the same as if we were still working in the fixed-projects-only reality. Historically, the outsourcing partner was hired for a set scope of work that had to be delivered in a specific time frame. The managers on the client's side used to communicate with a project or account manager from the outsourcing company who kept them up-to-date on the progress. We call it the client-manager communication model.

![Client-manager communication model](https://wise-site.fra1.digitaloceanspaces.com/communication_model_scheme1_1c979cf44a.svg)

 

## Client-team approach to software outsourcing

 

With new approaches to product launch, flexibility became the key success factor for most startups. By flexibility, we mean the ability to pivot the project on the go, adjust functionality based on the client’s expectations and feedback. For this, the old approaches to communication between the client and outsourcing team simply don’t work.

That’s why it is vital for the startups who [outsource MVP development](https://wise-engineering.com/expertise/mvp-development/) to be able to communicate with each member of the remote team individually. Because when an unexpected problem arises, people who are the closest to it can get the best solutions faster. That is what the client-manger model lacks – the ability to talk to the right specialist, the minute you need it. 

At Wise, we understood the problem of the communications via the manager long ago and rarely used it within our project. Today, what we advise our clients is to use client-team communication right from the start, no matter if it is a new product that we launch or scaling existing one with the new functionality. It does not mean the team operates without a project manager on our side (though sometimes it does), but that together with the PM the client can talk to the whole team or each specialist directly.

![Client-team communication model](https://wise-site.fra1.digitaloceanspaces.com/communication_model_scheme2_1bcfdcb7f4.svg)

 

## Two models compared

 

The infographic below illustrates how the business logic behind the startup development has changed. The key difference lies in the project planning and scope. You no longer need to know how to fully realize the functionality you have in mind, or have the comprehensive list of features you need to implement. Instead, you can communicate your business goals and key problems your product addresses, and together with your outsourcing partner decide on the best solutions to make those goals achievable. One condition only – to be able to create a product you’ve envisioned, engineers need to discuss it with you directly.

![Two models compared](https://wise-site.fra1.digitaloceanspaces.com/communication_model_infographics_6241028515.svg)

 

## Self-organized teams in outsourcing

 

Another phenomenon that was aroused by the change of the business development approaches, is the new paradigm for the outsourcing team organization and the rise of self-organized teams in the sphere. 

Usually, when this topic is discussed it is spoken of in the context of in-house teams. Though the covid pandemic has indisputably accelerated the remote-work trend and the way teams collaborate has been changed fundamentally. Self-organization became an even more critical aspect for all teams seeking efficiency. No matter if they work in-house or are part of the outsourcing project.

**What is a self-organized team?** It is a team whose members have the freedom to make decisions within their domain of expertise and not wait for their manager to assign tasks. A self-organized team has a clear vision or a product broader picture and does not expect detailed task descriptions in Jira. They communicate directly with the product creator and stakeholders and suggest solutions to the problem the management faces or users report. 

Simply put, this is a team that has the competence to accomplish the work without depending on people outside the team. Let’s discuss why this type of teamwork is beneficial both for the engineers and the business owners.

 

### From developers’ perspective

 

- In self-organized teams, people are mission-driven and understand the value of the product they build, as well as their own impact on its development. This also increases their level of involvement and promotes the ownership attitude.
- Engineers in self-organized teams tend to stay in the project longer as the process is dynamic and never becomes monotonous.
- When engineers take the responsibility for the decision they make, they are more eager to go the extra mile for their end-users. This helps them grow business thinking and changes the overall approach to product engineering. 
- Because each team member is involved in communication, the chances of spotting any miscommunication at the early stages are much higher. And it is always better to correct the concept of the feature implementation than to change the logic behind it entirely after it is almost ready.

 

### From the business owner’s perspective

 

- Working with a self-organized team you can finally minimize multi-tasking and micromanagement that eats up an enormous amount of time.
- With tech experts who handle the tech part, you can focus on the product strategy, hypothesis testing, and product growth that really deserve your rapt attention.
- To function effectively, self-organized teams should consist of experts in specific spheres. This also ensures all the tech-related decisions are made by people equipped with the best possible knowledge.
- With the clear communication loop established, all the people involved can take initiative and act faster. This helps to achieve frequent product delivery and fast product iterations.

 

## Wise approach to communication

 

Clearly, not every outsourcing team is ready to work within the self-organization paradigm, and the formation of such a team does not happen overnight. At Wise, we’ve been launching products for our clients for more than 15 years and noticed that to be able to effectively self-organize, a team should be working together for a significant amount of time. In our case, 60% of our teams work together for 3+ years and that’s the minimum time that is required. We also have teams that work together 5+ and 10+ years, and their achievements in self-organization are much more impressive.

Our project managers shared these handy tips on how to collaborate effectively with a self-organized team:

- **Minimize top-down management of tasks and activities.** Instead, you can build your cooperation model on partnership and strategic collaboration. For this, find a team that would consider your success their own achievement.
- **Do not confuse brief communication with clear communication.** Sometimes meetings last longer than usual. Sometimes an extra meeting is needed. The same as you now stay flexible with the project scope, adjust the communication schedule on the go. Our engineers usually request additional meetings if there are important topics to discuss.
- **Do systematic health checks to see if everyone from the team knows what to do and what is expected of them.** We usually do it with the *Roles & Responsibilities* exercise when each team member writes down what he/she expects of the teammates and sees what is expected of him/her. The points that do not coincide gets discussed.
- **Keep people in your outsourcing team informed.** The more the experts you’ve hired are up-to-date with the overall project situation, the more context they receive, the better solutions you’ll get on the outcome.
- **Clarity is the magic pill for effective communication.** Before you start moving forward with your outsourcing team, specify how you see the product development in the long perspective, how you measure success, what traction you expect, and what hypotheses you are trying to test. 

 

## Key takeaways

 

Software development outsourcing is a strategic initiative that allows startup companies to launch their product MVPs to market faster and validate ideas that will receive further investments. To succeed with the project you outsource, look for a custom software development company that’s open for dialogue, and would care about your product mission.

At Wise, we bring together engineers with deep expertise in [custom software development](https://wise-engineering.com/expertise/) and work in self-organizing teams. We’d be happy to discuss your business ideas and see how we can bring them in front of your end-users as fast as possible.

We create our story every moment