Implementing CRM and marketing automation: where to start and how to prepare the plan

Companies that start the journey by choosing a tool usually buy features they don’t need and miss the ones that would bring the most. This guide walks you through five steps: from the first internal conversation to a specification that vendors can answer with comparable proposals.

Where do you start when a company wants to implement CRM and marketing automation?

Start with the question of what should change in the business - not which tool to buy. The whole process runs like this:

1. Define the business goals. What exactly you want to improve, and by how much.
2. Map the sales and marketing process. Where opportunities, data and time get lost.
3. Set the scope and the order. CRM, marketing automation or both - and in what sequence.
4. Write the specification. A list of requirements by area that lets vendors submit comparable proposals.
5. Run the selection. Invite 2-4 vendors, score them against predefined criteria, and decide.

The first three steps are internal work. Nobody can sell them to you - which is exactly why most companies skip them. Then they wonder why they received five proposals that can’t be compared.

Step 1: define business goals, not technical wishes

The goal is not “implement a CRM”. The goal is a measurable business change - for example: shorten the sales cycle by 20%, lift enquiry-to-deal conversion by 15%, double the share of customers who make a repeat purchase, or cut manual campaign work in half.

Write every goal down with a number and a deadline. That list later becomes the basis for your questions to vendors and for judging whether the project succeeded. If you can’t measure a goal, it isn’t a goal - it’s a wish.

A practical test: for every requirement someone later proposes for the specification, ask which goal it serves. If there is no answer, it doesn’t belong in the project.

Step 2: map the sales and marketing process

Before you can automate anything, you need to know how the work runs today. Map both sides:

The sales side (CRM):

• Where do enquiries come from and who picks them up?
• Which stages does a sales opportunity pass through, and who decides on the transitions?
• Where do you lose the thread today: unanswered enquiries, forgotten follow-ups, data scattered across Excel and email?
• What reports does management need, and how much manual work do they take today?

The marketing side (marketing automation):

• Which customer segments do you address, and through which channels (email, SMS, web, social media)?
• Which communications repeat and could be taken over by automation: welcome series, abandoned carts, contract renewals, post-purchase sequences?
• What customer data do you actually hold, where is it stored, and do you have consent for direct communication?

The result of the mapping is a list of 5-10 specific use cases. A few real examples from our projects:

• a customer submits an enquiry on the website → a task opens for the salesperson in the CRM immediately, with all the customer’s details (this is how our own portal works too);
• a customer checks their credit limit with an instalment-financing provider but doesn’t buy → an automated campaign nudges them towards a first purchase;
• a customer has nearly repaid an existing loan → a campaign for a repeat purchase or for using the available limit;
• an existing bank customer with an active loan → an email with a personal calculation (their instalment, possible additional funds) instead of a generic push to a loan calculator (Addiko Bank);
• a customer has an open support ticket → the system can automatically exclude them from sales messages for that period (a case for why CRM and marketing automation must be connected);
• a customer abandons a cart → an automated prompt to complete the purchase;
• a customer shows signs of leaving → a win-back campaign, which at Shoppster won back 27% of customers who were about to leave.

These scenarios are the backbone of the specification - the vendor must show how they deliver them, not recite feature lists.

Step 3: CRM, marketing automation, or both at once?

It depends on where you lose the most. A rule of thumb from our experience:

• B2B companies with a sales team: CRM first. Without an orderly sales process and clean data, marketing automation has nothing to work with. Automation on messy data only sends the wrong things to the wrong people faster. We run our own business the same way: FrodX manages marketing, sales and customer support on its own HubSpot portal.
• B2C and retail: the platform for automating and personalising communication often comes first, because most of the value sits in post-purchase communication and customer retention. At Shoppster, automated tactics on one market delivered up to 48% of revenue from Emarsys activities, and the average order value was 14% higher.
• Both at once makes sense when sales and marketing are tightly intertwined - but split the project into phases with clear milestones, not one big bang.

Plan in phases: the foundation first (data, processes, basic usage), then quick wins (the 2-3 scenarios with the biggest impact), and only then expansion. The first phase should show a visible result within a few months, because a project without early results dies internally.

Step 4: how to prepare the specification and the RFP

The specification (and the RFP built on it - the request for proposal) should contain six elements:

1. a short introduction of your company and business context,
2. the main goals of the project (from step 1),
3. the list of requirements by use case (from step 2) - the core of the document,
4. your key specifics and challenges,
5. instructions on the timeline, format and submission of proposals,
6. legal provisions (confidentiality, NDA).

Group the requirements by area. For each area, here are examples of open questions for vendors:

Sales process and CRM: How does your solution support our sales process by stage? How is a lead handed over from marketing to sales and back? How does a salesperson work with the solution in the field, and how much data entry does each activity require?

Data and integrations: What integration options do you offer for connecting our existing systems (ERP, online store, contact centre)? How quickly is new data available for use in campaigns? How do you merge data from digital and physical channels?

Segmentation and personalisation: What segmentation options do you offer? How do you deliver personalised content across different channels? How does the solution recommend the next best product or content for an individual customer?

Campaign automation: Which triggers can start a campaign automatically? Can a running campaign be paused, edited and restarted without complications? What A/B-testing options does the platform offer?

Analytics and reporting: How does the solution connect campaign metrics to business outcomes such as revenue and ROI? Which standard reports are available, and what can be customised?

Security and compliance: Which GDPR-compliance mechanisms do you provide? Which security certificates do you hold (e.g. ISO 27001)? What is your SLA?

Implementation and support: Who performs the implementation and how long does it take? What training and support do users receive? What exactly is included in support - don’t assume your definition of support matches the vendor’s.

Three rules for writing the questions: ask open questions, not yes/no (the way a vendor answers reveals their experience); ask about business impact, not feature lists; and always require a demonstration and trial access. Don’t trust a vendor who avoids the demo.

And the most important rule of the whole document: the specification must be specific enough for the proposals to be comparable. If you receive proposals from €10,000 to €100,000 for the “same” project, you haven’t learned anything about the market - you’ve received proof that the brief was loose. Compare apples with apples.

Step 5: run the selection and make the decision

Invite 2-4 pre-screened vendors. Score their answers against criteria derived from your requirements - the scoring sheet later lets you justify the choice clearly to management as well. Before deciding, require a demonstration and trial access on your own scenarios from step 2, not on the vendor’s prepared examples.

And in the end, decide. A selection that fades out without a decision is the most expensive version: you paid for it with everyone’s time, and got nothing in return.

Who should take part in the preparation?

The specification is a team effort, not one department’s task. Involve: the head of marketing, the head of sales or e-commerce, the end users (salespeople and the campaign team - they will live with the tool), IT (integrations, security), finance (costs, licences, ROI), procurement and legal (contracts, GDPR). The project should have one owner with a mandate from management.

The department you leave out will show up mid-selection with its own requirements and stretch your timeline. Collect everyone’s requirements at the start, remove the duplicates, and turn them into scoring criteria.

How long do the selection and the implementation take?

A rough selection timeline: preparing the specification and RFP 2-4 weeks, vendor responses 3-4 weeks, review and scoring 2-3 weeks, presentations by shortlisted vendors 1 week, negotiations and the final choice 3-4 weeks. Around three months in total - and once you have estimated the timeline, double it, because companies almost never give the selection enough time.

The implementation itself is a separate project. Smaller B2B projects wrap up in 6-8 weeks from contract signature, the average across our projects is around 16 weeks, and larger B2C implementations can take six months or more.

If you are replacing an existing solution, negotiate an extension of your current contract in good time. Mid-selection, with a licence about to expire, you have no negotiating power.

The most common mistakes

• Choosing the tool before defining the goals. You are shopping for a solution to a problem you haven’t defined.
• A loose specification. The result is incomparable proposals and a choice made on price instead of value.
• Yes/no questions in the RFP. Every vendor answers yes to everything. You learn nothing.
• Features before business impact. You get a bare list of technical specifications, not a solution to your problem.
• Too many invited vendors. More than 4 proposals means a superficial review of all of them. Shortlist earlier instead.
• A selection without a decision. A surprising number of selections quietly die because the project loses priority. Keep the focus to the end.

What does this look like in practice?

B2C: Shoppster (United Group). E-commerce on two markets, one Emarsys account, nine or more automated, revenue-focused tactics: personalised prices and product recommendations in emails and on the website, personalised newsletters, win-back campaigns. The result: up to 9% of all e-commerce revenue attributed to Emarsys activities, 27% of at-risk customers won back, and a 14% higher average order value. Full Shoppster case study.

Personalisation in banking: Addiko Bank. Instead of a generic loan newsletter pointing to a calculator, an existing customer receives a personal financial scenario built from their own loan data: their current instalment and a calculation of how much extra funding they could get. We manage the campaigns in SAP Engagement Cloud under the Growth as a Service model, with part of our fee tied to the results. The personalised approach nearly doubled the click-to-open ratio (CTOR from 5.69% to 10.96%) and brought 270 requests for personal consultation and 79 signed loans; in the best scenario, 37.6% of requests converted into a loan. Full Addiko Bank case study.

Premium products on a global market: Pegasus Yachts. Sailing yachts priced at around €1 million, sold through a digitally built relationship - the buyers are known before the first handshake. Since the partnership began in late 2020, revenue has grown fivefold. Full Pegasus Yachts case study.

A connected system: Leanpay. An instalment-payment provider where campaigns run on SAP Engagement Cloud, customer support runs on HubSpot Service Hub with an Aircall call centre, and satisfaction is measured with InstantFeedback. A case of CRM, support and marketing automation working as one: a customer with an open support ticket can be excluded from sales messages for that period, and a poor satisfaction score triggers immediate follow-up. Full Leanpay case study.

B2B: FrodX. For B2B we don’t point to someone else’s success story - we use our own business as the case study, so the shoemaker’s children don’t go barefoot. We run marketing, sales and customer support on our own HubSpot portal: from the first website contact to the offer, the project and the support ticket. If you would like to see what a working B2B setup looks like in practice, co-founder Igor Pauletič will personally walk you through our CRM - book a slot.

Frequently asked questions

Should you implement CRM or marketing automation first?

B2B companies with a sales team should as a rule start with CRM, because automation needs orderly data and processes. B2C and retail often gain more from a platform for personalised communication. Base the decision on one question: where are you losing the most revenue today?

How many vendors should you invite?

2 to 4, pre-screened (references, a demo, independent reviews on portals such as G2 and Gartner). More proposals don’t mean a better decision - only more review work.

How much does implementing CRM and marketing automation cost?

Annual subscriptions range from a few thousand euros for basic tools to €100,000 and more for enterprise platforms. A typical B2B project in our practice lands between €20,000 and €50,000 (first-year subscriptions, implementation and user onboarding). B2C projects are larger, and the systems cost more because of the higher contact volumes - typically between €50,000 and €100,000.

What if we already have a CRM and are only adding marketing automation?

The steps stay the same, but the centre of gravity shifts to data: the quality of the data in your existing CRM, the consents you hold, and how the new tool will integrate. Poor CRM data is the most common reason marketing automation fails to deliver results.

How long does the implementation take?

Smaller B2B projects are completed in 6-8 weeks from contract signature, the average across our projects is around 16 weeks, and larger B2C implementations take six months or more. The vendor selection before that takes roughly three months.

Do we need an external consultant?

Not necessarily for steps 1-3, if you have the internal knowledge and the time. The specification and running the selection, however, are jobs where experience sharply reduces the risk - the wrong platform choice is a cost that drags on for years. It already helps to have someone who has seen many such projects review your draft specification.